A.I. - What You Need To Know!

A.I., OR A.G.I., OR  A.S.I.?

I have resisted commenting on the advances of Artificial Intelligence technologies. Recent news announcements have made further delay impossible. For those not sure about all this alphabet soup, allow me to give you a quick definition of these three seemingly mystical terms.

“A.I.” - Meaning “Artificial Intelligence”, is the subdivision of computer science which creates systems that perform tasks that normally require human intelligence. These tasks can include basic problem-solving, the appearance of learning, the purported understanding of language and response, the recognition of patterns, and the making of preprogramed decisions. It works because of massive computers that crunch human written programs called algorithms.

“A.G.I.” - Meaning “Artificial General Intelligence”, creates systems that appear to have the same cognitive abilities as a human with a very wide range of skills, talents, and knowledge. It appears to the person interacting with it to be able to reason, plan, and learn in such a way as to be indistinguishable from interacting with another human being. While it does not have emotions or creativity it appears to mimic them by analyzing millions of personal profiles to give a “sensitive” human-like response to the user. While once thought to take another five to ten years to develop, recent advances in computer chips, massive economic investments, and new methods of programming will make achieving AGI happen this year, by the end of 2025.

“A.S.I.” - Meaning “Artificial Super Intelligence”, creates systems that give the appearance of god-like intelligence. It would appear to be much smarter than humans in every possible way. Imagine a machine that can think, learn, and solve problems better than the smartest people on Earth - not just in math or science, but in the appearance of creativity, emotions, and even social skills.

If today’s AI were likened to a toddler, and AGI to an adult, then ASI would be like a super-genius who knows everything, learns instantly, and can improve itself at an unimaginable speed. In short, ASI would be an intelligence far beyond a single human’s capability, yet created by humans writing the programs, at least initially.

I would suggest that you read books like the following, or watch the movie adaptations of them if available. Books by William Hertling, especially “A.I. Apocalypse” and the book or movie “I, Robot” by Dr. Issac Asimov. Be sure to read about the Three Laws of Robotics and their failure.

At an ElevenLabs London Hackathon, two A.I. bots were set up so that one telephoned the other. At first, the two conserved in English, however, when “they” recognized each other as bots, they changed from English to what is now termed, “Gibberlink mode” (it sounds like the old dial-up modem screeches), in order to speed up their communication. No human can hear it and understand it. Please look up this link on YouTube:

https://youtu.be/EtNagNezo8w

Gibberlink was created by two META (Facebook, Mark Zuckerberg) engineers, Boris Starkov and Anton Pidkuiko. It uses a protocol called “GGwaves” that transmits data over sound instead of full words. According to the developers, it is 80% more efficient than using spoken words.

This event was foreseeable. The same basic experiment happened at Facebook back in 2017, when two A.I. programs seemed to create their own way of communicating and the engineers couldn’t understand the communication. Their answer back then was to simply turn them off.

These two events beg the question: If two or more bots can communicate without human knowledge how can we set up ethical and security safeguards? The answer is: we can’t.

Technology is neither good nor evil; it is what sinful people with an Adamic nature do with it that makes it good or evil. The biggest threat to mankind would be the weaponization of A.I. into a Frankenstein monster that we can no longer control. Giving A.I. its own secret language is only inviting the use of its evil side.

The good news is that A.I. is not creative. No one, not even human programmers, can create anything greater than themselves. Neither humans nor A.I. programs know how to make themselves “better”. All they can do is write bigger and bigger programs to mimic intelligence, but it is still ARTIFICIAL.

Various Big Tech firms have already programmed A.I. algorithms that have passed the Turing Test(!), named after famed computer pioneer Alan Turing. His life and achievements in breaking the WW II German Enigma Machine codes were masterfully recounted in the 2014 movie The Imitation Game.

In 1950, Turing proposed a test to determine if a machine program had reached the point where it had demonstrated human-like intelligence. In the test, if a human is in a written conversation with two entities, a person and a computer, and cannot distinguish between the two based on language usage, the machine has past the Turing Test.

The machine does not have to be correct nor logical; it only has to be convincing in simulating human conversation. In essence, the machine only has to be deceptive enough to fool someone into believing that there is a human “at the other end” of the conversation.

The criteria to pass the Turing test are:

1. Natural Language Processing (NLP): The machine must understand and generate human language fluently.

2. Knowledge Representation: The machine needs to handle and manipulate knowledge to provide contextually relevant responses.

3. Reasoning: The machine should demonstrate some form of logical reasoning, even if flawed, to sustain a conversation.

4. Learning: Ideally, the machine should learn from the interaction, adapting its responses over time.

Written programs have passed this test to appear human-like to humans. This ability is similar to The Tower of Babel, when the whole world spoke one language and nothing that was humanly possible was impossible for them. I know what God did in response to that development and now we will see what He does with this one.

Everyone seems to be talking about the incredible amount of electric power that the “A.I. Age” is going to consume. I have heard several knowledgeable people commenting that we are already using three per cent of our national power grid to power the A.I. server farms currently in place. At the same time the Environmental Terrorists want us to be forced into using all electric vehicles, including the nonsense of using battery powered military tanks, semi-tractor trailers, and farm equipment.

The futurists declare that the USA will need 99% of our current power grid to power all the new A.I. server farms that will be installed by 2027. That goal would require a phenomenal amount of new electric power generation facilities to be built and to be built very quickly. This need is not something that the current governmental structure could license and get in place in time.

People wonder why the Chinese Communist Party has built hundreds of dirty, coal fired, high pollution electric power plants - for which they have no fuel - all around their country. The reason is simple; they know that they will need that power if they are to successfully fight the upcoming A.I. wars. They will be ready except for getting enough coal to burn in them.

We should all be concerned at the prospect of all this needed new power plant construction. In order to achieve the required power production, the USA has committed to having private industry licensed to build their own nuclear power plants. The power generation necessary for this brave new world will require one nuclear power plant per server farm.

So, we need to build an electric grid large enough to charge battery powered vehicles and to win the A.I. war that is coming.

What I don’t hear anyone talking about is the almost equal amount of power that we will have to provide for the Blockchain technology that will parallel the A.I. advancements that are coming very, very soon.

 Blockchain technology was invented in 1991 to set the highest level of security for digital transactions. The development of cyber securities, like Bitcoin and Ethereum cryptocurrencies, required the development of a system in which no one could alter the timestamps of records - basically Blockchain is an unhackable security system.

At the heart of the Blockchain technology it is a decentralized business ledger that records transactions from more than one computer, encrypts it, checks and rechecks it across multiple nodes, making the information immutable. The energy consumption to make this work will rival that of the companion A.I. server farms!

What’s AI done to/for/against you in 2025? Here is a list of just ten items.

1) AI helped design a new breast cancer treatment with zero cancer drugs. Using high-cholesterol and anti-alcohol meds, ChatGPT-4 suggested combos, and some worked better than standard care.

2) AI is cloning voices by grabbing a clip of your child’s, spouse’s, boss’ voice from social media and podcasts. Then it calls your mom, your grandpa, your spouse saying that they are in big trouble and need money. Almost 72% of scams start with personal data grabbed online. This scam works because it feels real.

3) Chatbots are warping reality and folks are forming deep emotional attachments to a virtual AI “person”, even “creating” their desired perfect mate. It’s ruining mental health, causing depression, and influencing some people to commit suicide.

4) AI can predict massive wildfires 10 days in advance by using CO₂ data, land maps, and weather info with a better track record than professional weathermen.

5) Chipotle is opening a new store every week. They will use their “Ava Cado” chat bot to interview candidates, collect info, and send out offers. It cuts hiring time by around 75%.

6) In Israel, ChatGPT-4 is making 100% accurate decisions with cardiac patients on whether to implant stents or perform heart transplants, and decides in one-third the time of human doctors.

7) A biotech company is developing a home test to detect cancer using specially trained dogs and AI. You breathe into a mask for three minutes, mail it back, and the pups sniff for signs of cancer while AI tracks their reactions. Clinical trials showed 94% accuracy.

8) ChatGPT can plan a week of simple, healthy meals that you can actually cook, and make a grocery list with the exact ingredients. You can take it a step further by also asking it to stay within your budget or avoid foods you don’t like.

9) With the help of AI, sensors, and GPS we have developed the world’s first supersonic artillery shell (Mach 3.5 or 2,687 mph), the Sceptre TRBM 155HG. It can hit targets up to 100 miles away with an accuracy of 15 feet. That puts it well outside the reach of most jamming systems.

10) Klarna, a major financial tech company, just replaced 700 customer service people with one chatbot, powered by OpenAI.

We could lose up to half of all entry-level white-collar jobs within three years. However, economists argue AI will augment jobs, not totally destroy them. History shows us that automation leads to different jobs, not zero jobs. ATMs did not replace all bank tellers. In the future, we are not going to be adjusting our workforce over decades; we are going to be adjusting them in fiscal quarters.

AI isn’t just the future. It’s now. Here’s the best part: You don’t need a computer science degree to get going. Anyone can learn the basics. At least 80% of businesses will use AI by 2030.  Learning how AI works will make you more valuable in your job and help you think more critically about the tech you use.

There are free courses available for those who want to learn more, especially if you are job hunting and a knowledge of AI is a requirement for the open position. Below is a list of good, easy, free courses:

1. Elements of AI.           
Offered by the U. of Helsinki, designed for people who don’t have a tech background.      
elementsofai.com

2. IBM Skills Build           
IBM is giving away tons of free AI, cybersecurity, and data courses through this platform.        skillsbuild.org

3. Coursera – AI For Everyone   
One of the most popular beginner AI classes. coursera.org/learn/ai-for-everyone

4. Harvard’s CS50 AI       
For the more advanced. cs50.harvard.edu/ai

Pick one of the above courses and give it just 20 minutes a day. Then, you’ll understand what’s under the hood of today’s smartest tech.

The good news is that there are good paying jobs that AI cannot do: accountants, CEOs, forest fire fighters, flight attendants, electricians, plumbers and pipe fitters, chefs and head cooks, hearing aid specialists, security and fire alarm installers, machinery mechanics, hotel managers, aircraft cargo supervisors, insurance sales pros, and more.

It is important to differentiate between robotics and AI programs. Most people who are afraid of AI are actually afraid of robotics. Perhaps they have watched too many movies like Arnold’s “Terminator” movies, or “I, Robot”.

Robotics are very beneficial for doing repetitive, boring, tedious, and perhaps dangerous work. The problem isn’t the robots, the problem, as with AI, is the humans that are programming them!

A.I. has incredible potential in the areas of health care, basic education, agriculture, transportation, finance, and the environment. It will provide productive tools for countries, businesses, and individuals at a fraction of the current costs. Your use should be for questions like: “How can I make this document better?” Or, after plugging in a list of numbers or any other data: “Analyze this information. Explain any trends or insights you find.”

On the other hand, it will give cutting-edge power for enemies like China and North Korea to develop new debilitating cyberattacks against civilian and military targets. It will foster state-sponsored theft of intellectual properties. Thieves are using it to steal untold billions of dollars.

Never use it as a substitute for social contact versus having a real human interaction because if you do, it will create a form of emotional and social poverty = people marrying virtual spouses.

Above all, when thinking about AI, remember that it has: NO moral compass; NO common sense; NO intuition; NO wisdom; NO imagination; NO creativity (it only manipulates what already exists); and, NO critical thinking ability.

There are many AI programs; some for general purposes, some for specific purposes, and some for use in very specialized areas of knowledge. The programs are not interchangeable and each one will give different results (garbage in - garbage out). Each program will give one answer per query, but they will not give you pages of informational sites like those of traditional Internet search engines (although they, too, are flawed by the bias of their corporate leadership). If you use it, you should use it as your first step, not your last.

After you get a result, you should test the result of any AI search by asking: “How do you know?” You are asking the program what are the original sources of its’ information or is it just guessing. AI is quite capable of making things up, even if it is just trying to please you.

These programs will lie to you! AI is based upon gathering patterns and looking at trends. AI uses “Recursive Self Improvement” or “Reinforcement Learning with Human Feedback” in order to train the models. This leads to sycophancy, the offering of responses that please the user at the expense of being truthful.

Recommendations: Create a separate email address just for using AI. Use AI on a separate device to your regular computer, like an old phone or computer that you don’t use for anything else. Treat everything like public information. Don’t enter anything personal or sensitive to your life or business.

AI algorithms are written by human beings born with an Adamic Sinful Nature. Therefore, AI algorithms are as flawed as the people who wrote them. Approach using them with that attitude and you probably won’t go wrong.